Welcome to DCD, home of the number one construction magazine!
Welcome to DCD.com!

 Current Issue
 Click here to
 read the issue.
Click Here To Access The DCD Archives™
Subscriber Login

   Current Issue
   Issue Archive
   Specifiers Spotlights
   Building Products Revue
   Technical Articles
   Case Studies
   DCD Sq. Ft. Cost Guides

   Cost Trends

   Media Kit

   Free Subscription
   DCD E-News Subscription

D4COST Software

Optimization*[my building]
By Andy O'Nan

Is architecture an art or science? Is conceptual estimating an art or science? Actually, each is both but far more science than art and becoming even more so. If you don’t like the idea of changing from an artist into a scientist, you have advances in technology to blame. Modern information technology is providing access to much more data than ever which is pushing out the guesswork. The days of relying on people’s experience and intuition alone are over. Like driving a car guided by a GPS, we still need human judgment to make course corrections, but we can now improve the efficiency of the trip by reviewing alternative routes and avoiding issues like traffic. In this article, I’d like to introduce you to a scientific approach that leverages technology to combine design exploration with cost estimating and value engineering in order to find more optimal building solutions.

First, let me explain the parallel coordinate plot in Figure 1. This chart may look like abstract art but in fact it represents the output from a design and cost exploration analysis. Illustrated here are thousands of design options color-coded based on estimated total cost. The results show us the correlation of different design variables on cost (this example shows total lifecycle cost including the initial construction cost). From left to right you will see the variables explored across the x axis. Furthest to the left is Number of Buildings which in this case was either 3 or 4. The next is Number of Floors ranging from 5-8 floors per building. These two variables accounted for 8 options – perhaps more design options than are explored from a cost perspective on many projects. Moving further to the right on the figure, the number of options explodes into thousands as we looked at different shapes and orientations on site. Changing the variables can affect the cost per square foot of gross building area based on what happens to skin ratios, vertical circulation, MEP needs, site needs and more. This particular study shows a difference in cost from the baseline design (pink line) to the most cost effective design (dark blue lines) of about $30 million or 15%.

Is this science fiction? Is it purely academic theory? No, in fact, this is a real world case study. One of the keys to making it work is reducing the cycle time necessary to iterate thru options and generate the desired information. In a conventional approach, design teams generate the ideas and come up with the initial solutions. Then, contractors figure out how to build it, determine cost, and occasionally are given the opportunity to enhance the design with their own ideas – particularly when the design is already over budget! This process is wasteful and takes too long to try and optimize. Even when early contractor involvement is utilized, the design space is largely unexplored due to time constraints. The chances are good that we are not even close to achieving an optimal solution.

But, first of all, what is optimal? Optimization often means reaching the project requirements at the lowest possible cost. The project requirements can also be translated into time, revenue and other values. Perhaps optimal is the highest lease revenue potential for a developer. Or perhaps optimal is the shortest construction time on a casino or lowest imbedded carbon on a LEED® certified corporate office. Optimal can be described as the best possible mix of desired function and cost.

To find the best mix, we need to determine and understand the project objectives, constraints and also what potential variables can be explored. Then, we need a process by which to rapidly test, report and evaluate a large number of alternatives. BIM technology is a crucial ingredient to this process because it can streamline the design-estimating cycle time. BIM can be used to automate quantity take off and can be linked with estimating software. Beck Technology uses a macro level BIM software that is conceptual and links cost information directly within the 3D model. This allows for quick manipulation of the graphical model thereby affecting the cost estimate in real time and in one place. Leveraging BIM technology for these rapid virtual prototypes can reduce iteration cycle time by a factor of 5-20x.

To achieve that level of productivity, pre-programing of a cost database is necessary so that the estimate will automatically adapt based on the state of the model. Ideally, this is done in such a way that there is no need to graphically model all of the details in order to get an accurate cost estimate. Each door with hardware does not need to be drawn to know that each room will require them along with install labor. Reverse engineering of historical projects to mine data is an effective way of developing the database and rules of thumb based on real world projects. This helps develop example interior departments, spaces, site components, cladding types and more that can accurately price based on their type, location, size and shape. In some cases, components like cladding are estimated based on percentages of each material relative to the model exterior surface area. This can be highly accurate and when estimated in this way, the model will produce different cost per square foot based on the resulting floor-to-skin ratio. The same concept can be applied to other areas of the project so that the square foot cost will vary for mechanical systems, vertical circulation systems and more. The cost database needs to be developed to perform on the specific building product type, region and design variables that the team expects to explore.

Once the database has been configured, the design and preconstruction team can create a baseline building solution with a full cost estimate. The model can then be manually modified in a rapid virtual prototyping process to create different alternative versions. Each version is saved and comparative data can be reviewed during or after the versioning process. This process is guided by the initial objectives, variables and constraints that were established, but the exploration may affect the direction such that new areas end up being explored that were not originally thought feasible or viable. The solution sets are narrowed as a team until paired down and a best compromise solution
is identified.

From a technology stand point, the next chapter is automating the generation of design and cost versions. In the testing lab, Beck Technology has been working on this for the past several years in order to leverage cloud computing to generate millions of versions in a fraction of the amount of time compared to the manual process. Parallel coordinate plots (such as figure 1) and scatter plots (such as figure 2) are examples of the reporting used to understand and correlate the huge amount of data.

Automated optimization techniques are clearly on the bleeding edge of what’s possible today and it very well may be years before it is used in the main stream. There are a number of innovative firms researching and developing the processes and technologies to enable true optimization and faster virtual prototyping which at some point will be highly disruptive to today’s process. It’s time to embrace your inner-scientist!

About the author: Andy O’Nan oversees Beck Technology’s global business development efforts. Prior to Beck Technology, Andy’s work in the concrete formwork industry included estimating, engineering, CAD drafting, field supervision and business development. Andy’s project experience includes a wide variety of types across the southwest US including water/waste water, transportation, commercial, residential, and industrial projects. At Beck Technology, Andy has lead the deployment of the company’s flagship BIM technology at over 120 firms including over 20 of the top 50 builders in the US. Joining Beck Technology in 2007, Andy currently directs the team responsible for all business development, customer support, and customer training. Andy is a graduate of Texas A&M University's school of engineering and is a frequent public speaker on AEC technology.

©2015 Copyright DC&D Technologies, Inc. All rights reserved. | DCD Construction Magazine | Email: webmaster@dcd.com